How smart am I test: the honest answer is a profile and a margin

0 The short answer

The honest answer to how smart you are is a profile of separate abilities, each placed on a percentile with a margin of error, and your own impression cannot supply it: a meta-analysis of 41 studies found that self-estimates and measured cognitive ability correlate at only about .33, a link that leaves a typical error near 14 IQ points. A normed test compares you with a described group of adults and reports how sure it is. ACIS is one concrete option: Quick costs 25 dollars for 6 subtests in about 45 minutes, Optimized 35 dollars for 13 subtests in about 110 minutes and Full Scale 59 dollars for all 20 subtests in about 175 minutes, prices read on November 4, 2026, with a free trial that needs no card. The decisive limit is that any online, unsupervised result is an estimate for personal use, not a clinical finding.

1 What are you really asking when you ask how smart you are?

The question how smart am I is several questions at once, and a test can answer only the ones that name an ability and a comparison group. When people type "how smart am I test", "am I smart test" or "how clever am I test", they usually mean one of four things. Some want to know whether they reason well with new problems. Some want to know how much they know and how well they use words. Some wonder whether they are quick, and some wonder whether they can hold a lot in mind at once. Everyday talk uses one word, smart, for all of it.

Psychologists separate these, and the separation is the first step toward a useful answer. In the Cattell-Horn-Carroll framework, described on the page on the CHC model of cognitive abilities, broad abilities include fluid reasoning (Gf), comprehension-knowledge (Gc), quantitative reasoning (Gq), visual processing (Gv), working memory capacity (Gwm) and processing speed (Gs). A person can sit high on some and middle on others, which is why a single label such as "smart" hides most of what a careful measurement would show. The page on what IQ tests do and do not measure sets out the limits of the construct, including what no score captures, such as creativity in a studio or judgment in a crisis.

The second half of the question is "compared with whom". An IQ score is a position in a reference group, and a position needs a described group. If someone says "I am smart", the unspoken comparison is usually friends, classmates or colleagues, which are small, selected and not representative of adults at large. A normed test replaces that private yardstick with a stated one. The page on the six cognitive domains in IQ testing lists the abilities in plain language, and the rest of this page uses them.

Three honest points about the limits of this page. It does not teach you to estimate your own IQ, which another ACIS page covers in depth. It does not list the everyday signs people associate with high intelligence, and it does not walk through choosing among tests for a specific goal. It answers a narrower question: if you really want to know how smart you are, what counts as a trustworthy answer, what does it look like, and what should you do with it. Where a neighbouring page owns a topic, this page links to it instead of repeating it.

A last note on tone. Nothing in this page treats a score as a measure of a person's worth. A cognitive profile describes performance on a set of timed or untimed tasks on one occasion, compared with a group. It says nothing about kindness, character, effort or what a person can learn over a lifetime, and the group findings cited below are about averages across many people, never about you in particular.

2 How well do people judge how smart they are?

People judge their own intelligence only moderately well: across a large meta-analysis, self-estimates and tested ability share about 11 percent of their variance, so a confident guess is weak evidence. The strongest single source on the question is the meta-analysis by Freund and Kasten, published in Psychological Bulletin in 2012. It combined 154 effect sizes from 41 published studies and estimated the overall relationship between self-estimated and psychometrically measured cognitive ability at r equals .33. The authors describe the effect sizes as ranging from r of minus .25 to .85, with 144 of the 154 positive, and an accumulated sample size of 22,256 across those effect sizes.

What does r of .33 mean in practice? The authors put it in the units that matter here. Using the correlation to predict tested ability from a self-estimate, and assuming a standard deviation of 15 as for IQ scores, they report a standard error of estimate of 14.16 IQ points. They compare that with the standard error of measurement of a standardized ability test with a reliability coefficient of .90, which is 4.74 points, about one third as large. They conclude that using self-estimates as stand-ins for standardized tests is a rather imprecise endeavour. Shared variance tells the same story: the authors give 10.89 percent for the overall relationship. This is where our title comes from, and it is the authors' own figure.

Our arithmetic extends their number in a way worth seeing. A typical error of about 14 points, applied as a 95 percent band, would run roughly 28 points either side of a guess, which spans more than the difference between average and the top few percent. That is our own calculation on their published figure, not a result they report, and it assumes the simple prediction model they describe. The practical point survives any model: your impression of your intelligence cannot tell neighbouring percentiles apart.

The meta-analysis also found that accuracy is not fixed. Validity was higher when people rated themselves on a relative scale with clearly specified comparison groups, and when the ability rated was numerical rather than general cognitive ability. The authors state that the best conditions in their data produced a correlation of .78, and they note that most self-estimates, about 73 percent, came from a single question. In other words, a better question helps, which is part of why a normed test asks many.

An earlier, more specific line of work points the same way. Paulhus, Lysy and Yik tested whether self-report measures of intelligence could serve as proxy IQ tests in a study in the Journal of Personality. Using two undergraduate samples of 310 and 326 students who also took an IQ test, they found single-item self-reports correlated only about .20 to .25 with IQ, and even with indirect measures, aggregation and item weighting, validities did not exceed .30. They concluded that validities that stay near .30 undermine the usefulness of self-report as an IQ proxy in competitive college samples.

Two caveats keep this fair. The studies pooled by Freund and Kasten were mostly from Western countries and often used student samples, as the authors themselves note, so the exact figure may differ elsewhere. And a modest correlation is not a verdict that self-knowledge is worthless: the authors argue that the gap between what a person believes and what a test shows can itself be informative. The ACIS page on whether you can estimate your own IQ goes deeper into the self-estimation literature, including newer pooled estimates, so this page does not repeat it.

3 Does the Dunning-Kruger effect mean low scorers think they are geniuses?

No: the famous pattern is real as a description of averages in some data, but the strong reading, that the least able are the most overconfident and the most able are modest, is contested, and it is not a rule about any one person. The idea comes from Kruger and Dunning's 1999 paper, Unskilled and unaware of it, published in the Journal of Personality and Social Psychology. Its hypothesis, as later restated by critics, is that the degree to which people estimate their ability accurately depends on their ability. The paper is the origin of the popular line that incompetent people do not know they are incompetent.

The statistical critique matters for anyone reading a viral version of the claim. Gignac and Zajenkowski argued in Intelligence in 2020 that the standard test of the hypothesis, comparing self-assessed and objective scores across ability quartiles, is confounded by two general effects: people on average rate themselves above their peers, and extreme scores in one measure tend to be less extreme in another. Their title says the effect is mostly a statistical artefact. In a community sample of 929 people who rated their own intelligence and completed a matrix reasoning test, they reported no significant heteroscedasticity and an essentially linear relationship between objective and self-assessed intelligence, which is not the pattern the strong reading predicts.

The second effect they point to has its own meta-analysis. Zell, Strickhouser, Sedikides and Alicke reviewed the better-than-average effect in Psychological Bulletin in 2020, covering 124 published articles, 291 independent samples and more than 950,000 participants. They report that the effect is robust, with an effect size of 0.78 (95 percent confidence interval 0.71 to 0.84), and larger for personality traits than for abilities. In plain terms, most groups of people rate themselves above their group's average, which cannot be true for everyone. A "how smart am I" impression is therefore tilted by a general tendency before any individual difference enters.

What does this mean for you? Three things. First, if you suspect you are overestimating yourself, you are in the majority and your suspicion is not a diagnosis. Second, if you suspect you are underestimating yourself, the same literature gives you no reason to trust that suspicion either, because the correlation is moderate in both directions. Third, the useful step is the same either way: replace the impression with a comparison against a described group. The ACIS page on common myths about IQ tests covers related misreadings of what a score means.

The research here is about groups. None of it says that you are in any bracket, and none of it says that a person who feels unsure of their intelligence is probably wrong, or that a confident person is probably right. It says only that the feeling alone cannot be the measurement.

4 What can a viral smartness quiz tell you, and what can it not?

A quiz can entertain you and can even correlate with ability, but it cannot tell you how smart you are unless it names its comparison group, samples enough tasks to be precise and states its margin of error. Those three conditions come straight from the professional standards for testing, and a score that fails them has no stable meaning. The standard that governs this is the Standards for Educational and Psychological Testing, the joint publication of the American Educational Research Association, the American Psychological Association and the National Council on Measurement in Education, whose 2014 open access edition we read on November 4, 2026.

The first condition is a described reference group. Standard 5.8 says that norms, if used, should refer to clearly described populations, and that these should include the individuals or groups with whom test users will ordinarily wish to compare their examinees. A quiz that returns "your IQ is 128" without saying whom you were compared with has given you a number without a meaning. The group could be adults in general, or the quiz's own recent visitors, or nobody at all, and each gives a different score for the same performance.

The second condition is the number of tasks. The Standards state that if an assessment is shortened, for example by decreasing the number of items or tasks, reliability is likely to decrease, and that lengthening it with comparable tasks is likely to increase it. A ten question quiz and a 20 subtest battery are therefore not two versions of the same answer. The first is a rough impression and the second is a measurement with a narrower band, and the page on reliability and validity in IQ testing explains those terms with examples.

The third condition is stated uncertainty. Standard 2.3 asks that for each total score, subscore or combination of scores that is to be interpreted, estimates of relevant indices of reliability and precision be reported. Many quizzes print a single number with no band, which invites you to read it as exact. A trustworthy answer prints how far the number could move from one occasion to another.

Beyond those three, a short checklist helps before you share or believe a result.

• Does the page say who the norms are, how old they are and how many people they include?
• Does it report more than one ability, or only one total?
• Does it say how long the test is and whether the time is limited?
• Does it explain what the result should not be used for?
• Does it ask you to pay after the result is calculated, or to share it before you see it?

None of these checks proves a test is good, and a test can pass them and still be poor. They are quick ways to separate a measurement from a toy. The ACIS page on free IQ tests and what they give you applies the same thinking to the free options, and the sibling page on taking an IQ test without paying covers what a free score buys. For a ranked comparison of the products themselves, see the page on the best online IQ tests.

The point is not that a quiz is wrong to exist. A short quiz can be a harmless prompt for curiosity, and some free tools say plainly what they are. The point is that a feeling of having been measured is not a measurement, and the way the question is asked decides whether the answer can be trusted.

5 Which options can really answer the question, and what does each cost?

Only two kinds of option compare you with a described adult group on several abilities: an individually administered battery given by a psychologist, and an online battery with stated norms; everything else is a screening or a game. The table below lays them side by side. Clinic prices and the Pearson description come from pages we opened on November 4, 2026. ACIS prices were read on November 4, 2026, and can change, so check the current figure before you decide.

The psychologist row rests on two kinds of source. Pearson's product page for the WAIS-5 calls it an individually administered clinical instrument, gives its age range as 16 years 0 months to 90 years 11 months, and lists administration times of 45 minutes for the 7 subtest Full Scale IQ and 60 minutes for the 10 primary index subtests. The fees come from clinics, not from Pearson. Emory University's Psychological Center lists a comprehensive cognitive assessment at 900 to 1,800 dollars on a sliding scale based on income, with assessments running over several sessions of 2 to 4 hours each. The University of Colorado clinic called CU PATH lists a comprehensive adult evaluation at 2,100 dollars; that fee covers a broader evaluation, not an IQ test alone, and the clinic describes its work in terms of ADHD, learning disabilities and neuropsychological questions.

Read the table for what it does and does not say. A clinic assessment is the only option here that includes a trained person watching you work and writing an interpretation. It also costs far more and may involve a waiting list. An online battery is far cheaper and can be taken at home, at the price of having no observer. The ACIS page on where to take an IQ test, online and in person lists every route, and the sibling page on which IQ test to take helps match the route to your goal.

If your question is purely personal, such as wanting a defensible picture of where you stand before deciding anything else, the second kind of option is usually enough. If you need a result for a diagnosis, a school accommodation or a job, a clinic is the route, and the receiving party decides what it accepts. The sibling page on what an intelligence test is and how norms separate it from a quiz draws the same line from the other side.

6 Smart at what? The profile behind a single answer

A useful answer to how smart you are names the ability, because the same person can be strong in one broad ability and middling in another, and a single total hides that. Modern batteries are organised by the Cattell-Horn-Carroll model, which McGrew's overview, published in Intelligence in 2009, describes within the human cognitive abilities project. ACIS follows it with 20 subtests grouped into six primary indices, and its subtest pages describe each task. Here is how the six map to the question people actually ask.

• Verbal comprehension (VCI, Gc). How well you know and use words and knowledge. It includes Antonyms, Vocabulary, Information, Synonyms and Similarities. On the Vocabulary subtest, the page describes a 45 item task in which you type the meaning of each word, with 1 minute 30 seconds per item and no outside aids.
• Fluid reasoning (FRI, Gf). How well you solve problems you have not met before. It includes Matrix Reasoning, Figure Weights, Visual Number Series, Logic Grid and Complex Relations. The Matrix Reasoning subtest is described as a nonverbal task of 32 visual matrices, each missing a piece, timed at 30 seconds per item with a shared extra time bank.
• Quantitative reasoning (QRI, Gq). How well you work with number relations: Mathematical Achievement and Arithmetic.
• Visual spatial (VSI, Gv). Visual Puzzles, Layer Rotation and Spatial Comprehension. The manual says Layer Rotation carries no words at all.
• Working memory (WMI, Gwm). How well you hold and manipulate information in mind: Digit Span, Alphanumeric Sequencing and Visual Sequence. The Digit Span subtest page describes three parts, forward, backward and sequence, with digits played aloud at one per second.
• Processing speed (PSI, Gs). How quickly and accurately you scan and decide: Symbol Search and Coding. The Symbol Search subtest page describes a rapid visual comparison task of 80 puzzles with a 2 minute limit.

Why does a profile beat a total for this question? Because a total is an average, and averages hide differences. Two people with the same Full Scale IQ can have very different strengths, one verbal and one visual, and what each does well in everyday life may differ accordingly. The ACIS page on an IQ profile test and when a weakness is just noise explains how to tell a real difference from a wobble, and the page on index score differences covers how common gaps between indices are.

The Standards give a necessary warning about profiles. Standard 1.14 says that when interpretation of subscores, score differences or profiles is suggested, the rationale and relevant evidence in support of that interpretation should be provided. Standard 2.3 adds that reliability should be reported for each score that is interpreted, and its comment notes that a total score can be acceptably consistent while subscores have unacceptably low reliability. In plain terms, a single subtest score is a noisier reading than the index it belongs to, and an index is noisier than the total. Read profile differences as leads to consider, not as verdicts.

Abilities are also related to one another, not independent. The sibling page on the g factor and how it is tested explains how a general factor and separate abilities are handled together, and it is the place to go if you want the theory behind why a total still carries information.

7 How do you read the answer: scaled score, index, percentile and margin?

A result has four parts, and the one most readers skip, the margin, is the one that keeps the others honest. The first part is the score on its scale. On the standard IQ-style scale used for indices and the Full Scale IQ, the mean is 100 and the standard deviation 15; subtest results use scaled scores from 1 to 19 with a mean of 10 and a standard deviation of 3. The page on why the standard deviation of 15 matters explains the scale, and the page on how IQ scores are normed explains what a norm group is.

The second part is the percentile, the share of the comparison group scoring at or below you. On a normal curve a score maps to a percentile directly, and the table gives the common points. These are our arithmetic from the normal curve with a standard deviation of 15, not figures from any test.

The tail values answer a question many readers have, how rare a given score is, and the ACIS tool for how rare a given IQ is turns any score into a one in X figure. A related trap is averaging a percentile with a score, which the page on why an IQ score and a percentile cannot be averaged explains.

The third part is the margin. A test score has a standard error of measurement, and a confidence interval is built from it. Freund and Kasten give a worked case: a standardized test with a reliability of .90 has a standard error of measurement of 4.74 IQ points on a scale with a standard deviation of 15. Our arithmetic on that figure gives a 95 percent band of about 9 points either side. A reported 115 from such a test would then sit somewhere between about 106 and 124, which on the curve above is roughly the 66th to the 95th percentile. That is an illustration of a test with that reliability, not a statement about ACIS or about any particular test. The practical lesson is to read a score as a band, and every ACIS form reports its Full Scale IQ with a 95 percent confidence interval for that reason.

The fourth part is context: the date, the conditions, the form and the comparison group. The APA Guidelines for Psychological Assessment and Evaluation, approved in March 2020, ask psychologists to choose tools with sufficient validity evidence and score reliability (Guideline 6), to use multiple sources of information (Guideline 7) and to attend to the appropriate normative comparison and standardized administration (Guideline 8). Those guidelines are written for psychologists, not for people taking a test at home, but they show what a careful reading looks like: one score, one band, one occasion, read alongside other evidence about you.

8 Why is a battery more precise than one quiz or one subtest?

A battery is more precise because every extra comparable task is another independent chance for chance to cancel out, and the Standards state the principle directly: shortening an assessment is likely to lower reliability, and lengthening it is likely to raise it. The sentence in the 2014 Standards for Educational and Psychological Testing adds that lengthening the assessment, and thereby increasing the size of the sample of tasks, is an effective and commonly used method for improving reliability and precision. A quiz of a dozen questions samples a dozen tasks. A battery of 20 subtests samples many more, in different abilities, which is also why it can report a profile.

This is general psychometrics and it applies in both directions. A short test is not worthless, and a long one is not magic. The ACIS page on what a short battery costs you weighs the trade for the Quick form, and the page on IQ tests with instant results explains what speed of delivery can hide. A short form is a reasonable first look when you want an indication, and a longer form is the reasonable choice when you want to read separate abilities, because each index rests on more tasks.

One subtest deserves special caution for the question how smart am I. A single subtest result is a reading on one narrow task. A weak result may reflect a bad hour, a distraction, an unfamiliar format or a real difficulty, and a strong result may reflect a lucky guess pattern or a practised format. The Standards say in a comment that test manuals and score reports should discourage overinterpretation of information that may be subject to considerable error, and they single out subtest scores as a case where that matters. When you want to check one ability on its own, for example speed, a stand-alone option such as a processing speed test can give a quick look, but it answers a smaller question than "how smart am I".

Retaking is the other precision trap. People who get a surprising result often want to take the same test again, and a second score from the same test is not an independent check. Scharfen, Peters and Holling pooled the evidence in a 2018 meta-analysis in Intelligence: 174 samples from 122 studies, with an overall sample size of 153,185. They found significant retest effects, meaning that scores rise because the test was repeated, and no further gains after the third administration. They also report that the size of the effect depends on the ability, the equivalence of the forms, the retest interval and the participant's age. For you, the implication is plain: a higher second score does not show that you became smarter, and a very different second score from a different test may reflect a different comparison group and not a change in you.

9 What conditions make an unsupervised answer worth having?

An unsupervised result is worth reading only if you create the conditions a supervised session would have created for you, because the Standards treat unproctored administration as a reason for caution. The 2014 Standards, in their discussion of fairness, note that unproctored administrations, where standardization cannot be ensured, could give an advantage to some test takers over others, and that in such situations results should be interpreted with caution. In a comment on validity evidence, they list the mode of administration, for example unproctored online testing compared with proctored on-site testing, among the conditions that may change how far earlier evidence applies. That is not a reason to avoid online testing. It is a reason to take it carefully and to read it with that caution.

What can you control? These are sensible preparations, not claims from a study.

• Choose a quiet place and an uninterrupted block. A phone call in the middle of a timed task changes the task.
• Plan for sound. Digit Span, Alphanumeric Sequencing and Arithmetic present their items as audio and need working sound. The Digit Span page recommends headphones or a quiet room, and the three audio subtests do not work without it.
• Leave outside aids out. The pages for Vocabulary and Matrix Reasoning state that no external aids are allowed. A calculator, a dictionary or a second screen would answer a different question.
• Use the breaks the format allows. ACIS allows breaks and saves progress between subtests, and you have 30 days to complete, so you do not need to push through when tired. Take the long form over more than one day if that is how you do your best work.
• Do not take it in a state you would not want measured. If you are ill, upset or exhausted, the result describes that day. The 30 day window exists so that you can choose another one.
• Read the instructions and the practice items. Understanding the task is part of the standardization, and unfamiliar formats are a source of avoidable error.

Language matters too. ACIS is English only, and the manual says verbal and knowledge tasks are affected by English proficiency and educational exposure. If English is not your first language, read your verbal results with that in mind and weight the nonverbal and speed indices more. The ACIS technical manual, linked below, is the place to read the documentation for how each domain is described.

10 What should you do with the answer once you have it?

Treat the result as a dated estimate of where you stand on several abilities in a comparison group, use it for personal decisions that can tolerate error, and take any question with consequences to a person who is qualified to answer it. That sentence is the whole advice, and the rest of this section unpacks it.

Start with the number and its band. Read the Full Scale IQ as a range and the percentile as a position, not a rank among your friends. Then look at the profile: which index is highest, which is lowest, and how big the gap is. A gap of a few points is within the noise the previous sections described. A large and consistent gap, one that shows in several subtests of the same domain, is worth a thought about how you work and learn. The page on what your IQ result means and how to choose a test walks through reading a result, and the profile page linked earlier covers when a difference is real.

Then use the picture for choices a personal estimate can support. People use it to see which kinds of work and study fit how they think, to settle a long running curiosity, to compare their own sense of themselves with a measured one, or to decide whether a more formal assessment would be worth the cost. The page on signs of high intelligence and what the research shows is the place to look at everyday behaviours, which are a different kind of evidence from a test result.

Be careful with three things. First, a group finding is not your profile: an average difference between groups in a study tells you nothing about where you fall. Second, a lower result in an area does not describe your worth, your effort or your future, and a profile of timed tasks is one thin slice of a person. Third, do not carry the number into a setting that has its own rules. The APA's 2020 guidelines are written for psychologists who conduct testing, assessment and evaluation, and they stress selecting tools with sufficient validity evidence, using multiple sources of information and attending to the appropriate normative comparison. A home result is one source among several and does not replace that process.

If you have a concern about attention, learning, memory or mood, or if you need documentation for school or work, see a licensed psychologist. An online test cannot diagnose or document anything, and a clinic can. If you simply want to know more about where your strengths lie, an online battery is a reasonable starting point, and the clinic route remains open afterward.

11 What does a normed ACIS result give you, and what does it cost?

ACIS gives you a profile in place of one number: a Full Scale IQ with its 95 percent confidence interval, six index scores, scaled scores for each subtest and percentiles, for a one time price and with a free trial that needs no card. The prices below were read on November 4, 2026, and can change again, so confirm them on the site before you pay.

Several features matter for the question on this page. Every form reports a Full Scale IQ with its 95 percent confidence interval. Seven subtests appear only in the Full Scale form: Information, Arithmetic, Mathematical Achievement, Visual Number Series, Complex Relations, Visual Puzzles and Visual Sequence. A Full Scale IQ from a Custom Test needs the Full Scale battery, and from six subtests upward the Full Scale form costs less than the same number of Custom subtests: six Custom subtests cost 60 dollars, and the Full Scale form costs 59. Payment is one time, with no subscription. You can take breaks, progress is saved between subtests, and you have 30 days to finish.

The free trial needs no card and is limited to one per person: the site refuses a second trial from the same email address or the same browser. Finished trial subtests are kept if you later unlock a form that includes them. That makes the trial a low cost way to see how the tasks feel, how the report presents a result and whether the format suits you, before spending anything. If you want to see a report first, the page of a worked IQ test report example shows what the sections contain, and the example links in the boxes on this page open sample dashboards.

The refund terms are specific and should be read before paying: "Full refund within 5 days if the report does not deliver the features described at checkout or a technical issue prevents you from accessing your results." This is a conditional refund, not an unconditional money back guarantee.

12 Where ACIS sits, and what a buyer should do

ACIS is a normed, adult compared, online battery that gives a profile and a band, and it is also an unsupervised, English only, nonclinical tool whose result is an estimate for personal use, not a document anyone else is obliged to accept. The honest summary has two halves, and both belong in the same paragraph. On the strength side, it compares you with adult norms for ages 16 to 90, reports six indices and a Full Scale IQ with a 95 percent confidence interval, and costs 25 to 59 dollars one time against clinic fees of 900 dollars or more on the pages read on November 4, 2026. On the limit side, nobody watches you work, the manual says verbal and knowledge tasks are affected by English proficiency and educational exposure, and its own research page states that the test "is not intended for clinical diagnosis, hiring decisions, accommodation determinations or admission to high IQ societies".

The standard that frames this is the Standards for Educational and Psychological Testing (American Educational Research Association, American Psychological Association and National Council on Measurement in Education, 2014). It asks that norms refer to clearly described populations, that reliability be reported for every score that is interpreted, that profile interpretations come with evidence, and that unproctored administrations be read with caution. The APA's 2020 Guidelines for Psychological Assessment and Evaluation add the point about multiple sources of information. Applied to ACIS, that means reading the Full Scale IQ with its band, reading the indices as leads, reading the whole as one source and keeping the documentation in view. The technical manual is where the documentation lives, and a buyer who wants the evidence should read it before paying.

What should a buyer do? A short sequence works for most readers.

1. Decide what you want to know. If it is a personal picture of your abilities, an online normed battery fits. If it is a diagnosis, an accommodation, a job or a society, go to the party that decides and ask what it accepts.
2. Try the free trial first. It costs nothing and needs no card, and it shows you the format and the report.
3. Choose the form that matches your question. Quick for a first look, Optimized for more domains, Full Scale for the whole profile, or a Custom Test for one domain.
4. Prepare the session. Quiet room, working sound, no outside aids, breaks as needed, and a day you would want measured.
5. Read the result as a band and a profile. Do not retake it to chase a number.

The next step is the free trial, and the sample report beside it shows what you would receive.

13 Sources Behind This Page

Every figure on this page that is not our own arithmetic comes from one of the sources below, and each is linked at the point where it is first used. Web pages were opened on November 4, 2026. Nothing about ACIS beyond its published product facts is stated, and competing online products are neither linked nor used as sources.

• American Educational Research Association, American Psychological Association and National Council on Measurement in Education. Standards for Educational and Psychological Testing. American Educational Research Association, 2014 (open access edition).
• American Psychological Association, APA Task Force on Psychological Assessment and Evaluation Guidelines. APA Guidelines for Psychological Assessment and Evaluation. American Psychological Association, approved March 2020.
• Freund P, Kasten N. How smart do you think you are? A meta-analysis on the validity of self-estimates of cognitive ability. Psychological Bulletin, 2012, volume 138, issue 2, pages 296 to 321.
• Gignac G, Zajenkowski M. The Dunning-Kruger effect is (mostly) a statistical artefact: Valid approaches to testing the hypothesis with individual differences data. Intelligence, 2020, volume 80, article 101449.
• Kruger J, Dunning D. Unskilled and unaware of it: How difficulties in recognizing one's own incompetence lead to inflated self-assessments. Journal of Personality and Social Psychology, 1999, volume 77, issue 6, pages 1121 to 1134.
• McGrew K. CHC theory and the human cognitive abilities project: Standing on the shoulders of the giants of psychometric intelligence research. Intelligence, 2009, volume 37, issue 1, pages 1 to 10.
• Paulhus D, Lysy D, Yik M. Self-report measures of intelligence: Are they useful as proxy IQ tests? Journal of Personality, 1998, volume 66, issue 4, pages 525 to 554.
• Scharfen J, Peters J, Holling H. Retest effects in cognitive ability tests: A meta-analysis. Intelligence, 2018, volume 67, pages 44 to 66.
• Zell E, Strickhouser J, Sedikides C, Alicke M. The better-than-average effect in comparative self-evaluation: A comprehensive review and meta-analysis. Psychological Bulletin, 2020, volume 146, issue 2, pages 118 to 149.
• Pearson. Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5) product page, page read on November 4, 2026.
• Emory University Psychological Center. Frequently asked questions: services, fees and assessment sessions, page read on November 4, 2026.
• University of Colorado, clinical psychology program. CU PATH clinic: fee for a comprehensive adult evaluation, page read on November 4, 2026.

14 Frequently Asked Questions